ChatGPT Ads Are Live in the UK: What Happens to Citations

By , Co-founder, GeoLinks · · 8 min read
A UK marketing manager reviewing a ChatGPT ad slot next to an organic answer on a laptop screen, natural daylight in a London office
A UK marketing manager reviewing a ChatGPT ad slot next to an organic answer on a laptop screen, natural daylight in a London office

ChatGPT ads went live for UK advertisers in June 2026, the first European market to get self-serve access. Over 2,000 brands now run them worldwide through Criteo. Nearly every article about this covers the ad. Almost none covers what sits above it: the answer itself, still generated the same organic way it always was. This explains the rollout, where the ad actually sits, and why the brand the model already recommends keeps a slot no advertiser can buy.

Key takeaways

  • UK advertisers got self-serve access to ChatGPT Ads Manager in June 2026, ahead of Brazil and Mexico, which remain in planned status as of late July.
  • Criteo’s brand count on ChatGPT ads doubled from roughly 1,000 in early May to over 2,000 by June, spanning apparel, home furnishings, electronics, automotive and beauty.
  • Ads appear as a sponsored link below the AI-generated answer. They do not replace the answer or influence which sources it cites.
  • Only Free and ChatGPT Go users see ads. Plus, Pro and Enterprise users, the higher-intent professional tier, see none.
  • Roughly 20% of ChatGPT queries carry direct commercial intent, and ad dismissal rates have fallen 50% since the February 2026 launch.

The rollout is further along than most UK coverage suggests

Most UK marketing coverage still frames ChatGPT ads as coming soon. That framing is out of date. OpenAI’s ad pilot launched in the US in February 2026 at a $60 CPM with a $200,000 minimum spend, both steep enough to keep it enterprise-only. That changed on 5 May 2026, when OpenAI opened a self-serve Ads Manager to every US business, dropped the minimum spend entirely and introduced CPC bidding alongside CPM. Pricing has fallen with the wider access: CPMs now clear as low as $25 in some categories, and per-click bids commonly start around $3 to $5.

The UK followed in June 2026, becoming the first European market with self-serve access. Japan and South Korea went live the same month. Brazil and Mexico were named as planned markets back in May and remain unconfirmed as of late July 2026.

A small UK business owner setting up a ChatGPT Ads Manager campaign on a laptop at a home office desk, morning light through a window
Self-serve access removed the $200,000 minimum spend that kept ChatGPT ads enterprise-only.
MarketAccessStatus as of late July 2026
United StatesSelf-serve, no minimum spendLive since 5 May 2026
Australia, Canada, New ZealandSelf-serveLive, pre-UK
United KingdomSelf-serveLive since June 2026, first in Europe
Japan, South KoreaSelf-serveLive since June 2026
Brazil, MexicoPlannedNot yet confirmed live

OpenAI took the rollout to Cannes Lions 2026, pitching ChatGPT users to marketers as “super intentional”: people who arrive already deep in a task rather than idly scrolling. The company is reportedly targeting $2.5 billion in advertising revenue for 2026. That is not a pilot anymore. It is a real UK media channel, and it is already running.

The ad sits below the answer, and the answer is still organic

Here is the distinction the coverage keeps skipping past. A ChatGPT ad is a clickable sponsored link placed underneath the AI-generated answer. It is a separate object, positioned after the model has already finished responding.

The answer itself, the part that names a brand, recommends a product or cites a source, is generated the same way it was before ads existed. Nothing in OpenAI’s ad documentation or in the Criteo integration suggests advertiser spend feeds back into which sources the model selects when it writes that answer. The two systems run side by side, not one inside the other.

That matters because most GEO conversations conflate “showing up in ChatGPT” with “buying a ChatGPT ad.” They are not the same purchase. One buys a labelled slot underneath the answer. The other is earned, the same way it always was: through the sources, mentions and citations that shape what the model already believes about a brand before anyone opens their wallet.

ChatGPT ad slotOrganic citation in the answer
Who controls itAdvertiser, via bid and budgetThe model, based on training data and retrieved sources
Where it sitsBelow the answerInside the answer
Who sees itFree and ChatGPT Go users onlyEvery user, including Plus, Pro and Enterprise
How you get itPay per click or per thousand impressionsEarn citations on pages the model already trusts
Trust signal to the readerLabelled as sponsoredPresented as the model’s own answer

Who actually sees these ads

The tier gating is the detail most coverage buries. Ads appear only for Free and ChatGPT Go subscribers. Plus, Pro and Enterprise users, the tiers most likely to include B2B buyers, researchers and professionals running higher-value queries, see no ads at all.

That means the audience most worth reaching with a paid slot is disproportionately consumer, price-sensitive and running lighter queries. The audience with genuine purchasing authority for higher-ticket B2B and professional services is precisely the group the ad product cannot reach. For those buyers, the organic answer is the only exposure a brand gets, sponsored or not.

Split-screen style photograph showing a young consumer browsing ChatGPT on a phone in a cafe in the evening, and a professional reviewing ChatGPT on a laptop in an office during the day
Ads reach Free and ChatGPT Go users. Plus, Pro and Enterprise users see none.

The organic recommendation is the slot money cannot buy

OpenAI’s own numbers make the case for organic work stronger, not weaker. Roughly 20% of ChatGPT queries carry direct commercial intent, so most conversations that touch a purchase decision are not ad-served at all. They resolve through whatever the model already recommends. Ad dismissal rates have fallen 50% since the February 2026 launch, which OpenAI reads as rising relevance, but it also means users are engaging with the format rather than reflexively skipping it. That raises the cost of getting the organic answer wrong next to a slot readers are actually looking at.

Across the accounts we track citation performance for, none has ever appeared inside a ChatGPT ad slot. Every mention logged in our monitoring has been organic: a source the model chose to cite, not a placement anyone paid for. That is the same mechanism behind the Garden Ornaments result we cite often, a 776% increase in organic traffic, from 727 to 6,370 monthly visits in seven months, driven entirely by earned placements on pages ChatGPT and Google already trusted. Ads did not exist in that channel yet when the growth happened, and the growth did not need them.

The prepare-before-launch checklist

UK brands weighing whether to test ChatGPT ads should fix the organic side first. A paid slot next to an answer that never mentions the brand wastes the click.

  1. Check whether ChatGPT already recommends you. Run your category’s real buyer queries and see who the model names first, with no ad involved.
  2. Fix the citation gap before the ad budget. If competitors get named and you do not, that gap costs more clicks than any bid strategy fixes.
  3. Target Free and Go-tier intent, not B2B research intent. The ad product cannot reach Plus, Pro or Enterprise users, so match campaign goals to that audience.
  4. Treat the ad as a supplement, not a replacement. It buys visibility for one query. Organic citations compound across every query the model answers about your category.
  5. Track both channels separately. Ad-driven clicks and organic-citation clicks should never blend into one number, so set up AI referral tracking in GA4 before the campaign starts.

Multi-engine reality check

This is a ChatGPT-specific product. Claude and Anthropic have not launched an advertising product inside Claude’s chat interface. Gemini carries ads inside Google’s wider ecosystem, including AI Overviews, but Google has not confirmed an equivalent sponsored-slot format inside the Gemini app itself. Perplexity has trialled sponsored questions but at nowhere near ChatGPT’s 2,000-brand scale, and its citation panel remains organic. Microsoft Copilot has not announced a comparable ad product as of late July 2026, though Microsoft’s wider ad network gives it an obvious path if it chooses. Grok carries no publicly confirmed advertising product tied to answers yet. Google’s AI Mode and AI Overviews already show ads separately from the generated answer box, the closest existing precedent for the ChatGPT model. Treat ChatGPT as the leading edge here, not yet the shared standard.

Frequently asked questions

Are ChatGPT ads live in the UK yet?

Yes. OpenAI opened self-serve ChatGPT ads to UK advertisers in June 2026, the first European market to get access.

How many brands are advertising on ChatGPT?

Over 2,000 brands run ads through Criteo’s integration as of June 2026, up from roughly 1,000 in early May.

Where do ChatGPT ads appear on the page?

As a clickable sponsored link placed below the AI-generated answer, not inside it or in place of it.

Which ChatGPT users actually see ads?

Only Free and ChatGPT Go tier users. Plus, Pro and Enterprise subscribers see no ads at all.

Does buying a ChatGPT ad affect whether a brand gets cited organically?

No. The ad slot and the answer are generated separately, so a paid placement does not influence which sources the model cites.

What percentage of ChatGPT queries have commercial intent?

About 20% of ChatGPT queries carry direct commercial intent, OpenAI told marketers at Cannes Lions 2026.

How should a UK brand prepare before running ChatGPT ads?

Fix organic citation gaps first. An ad next to an answer that never mentions your brand wastes the click.

Run a free AI citation check to see whether ChatGPT already recommends your brand before you spend a pound on the ad slot underneath it, or talk to us about ongoing citation work that earns the answer itself.